GHSA-7rqj-j65f-68wh

Summary

The default email-address normalizer used by the email/magic-link sign-in flow validates the address before applying Unicode normalization. An address can contain a Unicode character that is not an ASCII @ (U+0040) but canonicalizes to one under NFKC/NFKD normalization (the normalization commonly applied by mail libraries and services for internationalized email). Such an address passes the normalizer's single-@ check, but a downstream mail library that normalizes the string then sees two @ separators and may deliver the passwordless sign-in link to a different recipient than intended. This is an instance of validating before canonicalizing.

Am I affected?

You may be affected if all of the following hold:

• You use next-auth >= 4.0.0, < 4.24.14, or @auth/core >= 0.1.0, < 0.41.3.
• You have the email / magic-link (passwordless) provider enabled.
• You rely on the built-in default identifier normalizer (you have not supplied your own normalizeIdentifier).
• Your sendVerificationRequest implementation uses a mail library or delivery service that applies Unicode normalization to recipient addresses (most internationalized-email/SMTPUTF8-capable senders do).

You are not affected if you do not use the email provider, or if your normalizer/mailer rejects or canonicalizes non-ASCII addresses before they are validated.

Impact

• Account takeover: an attacker who knows a victim's email address can request a magic link that is delivered to an attacker-controlled mailbox, then use it to sign in as the victim.
• No victim interaction is required to misroute the link; the attacker initiates the flow.

Patched version

The fix applies Unicode (NFKC) normalization before the address is validated, so homoglyph separators are collapsed and rejected up front. Upgrade to the first release containing this fix (pending; this advisory will be updated with the exact patched version before publication). No application code changes are...